Nature Genetics interview

Nature Genetics has an interview with me on their blog: 

The discovery that NOD^k mice with the insHEL transgene develop diabetes is described as being serendipitous. What were your initial thoughts about this? 

At the time we first found that NOD^k.insHEL male mice developed diabetes I was actually working on immune defects in NOD mice, rather than beta cell defects. My first thoughts were that this was just another immune defect, with the immune system attacking the beta cells because they expressed the insHEL transgene. Since it fit our preconceived ideas we didn’t take too much notice, but just to be safe I set up a backcross to eradicate the adaptive immune system from the NOD^k.insHEL mice. It took a couple of years for the mice to breed and age, so I had almost forgotten about the finding when the first immune-deficient NOD^k.insHEL mice started to develop diabetes. At that point I was really startled – the cross should have eliminated diabetes if it was immune-mediated. I knew then that we were looking at some completely new biology – which took another 10 years to dissect! 

What advantages does your new mouse model bring to the field?

There are so many aspects to diabetes that it is often impossible to untangle the causes of disease. For example, one of the critical clinical developments in type 2 diabetes is the death of beta cells. It marks a shift from insulin-resistant diabetes (which is largely treatable), to insulin-deficient/insulin-resistant diabetes (for which there are no effective treatments). But why are the beta cells dying? From previously mouse models there were many reasonable hypotheses that were put forward – maybe it is the demand placed on the beta cell for extra insulin production, maybe it is a toxic effect of high blood glucose levels, maybe it is a side-effect of the high fat diet used to induce diabetes in the first place, or maybe it is immune-mediated. Our model has the advantage that it can strip away all of these interactions to observe the direct effects of forcing beta cells to produce too much protein – a process that results in beta cell failure. Looking forward, I see a major advantage in using this model to screen for drugs that stop the loss of beta cells in type 2 diabetes, which is really the key unmet medical need in diabetes treatment.

One of your interesting observations is the difference in diabetes incidence between the male and female mice, mediated by male sex hormones.  What parallels are there with humans and how might you use this model to explore this further?

So far we have only seen diabetes in male insHEL mice, despite challenging female insHEL mice with multiple strategies that promote diabetes in male mice (diet, autoimmunity, genetic background). At a cellular level it looks like male islets are just under more metabolic pressure than female islets, such that the insHEL stress is enough to make male mice diabetic, while female mice stay healthy. This could actually explain a lot about the epidemiology of type 1 diabetes in humans. Most autoimmune diseases have a strong female bias, while type 1 diabetes has a weak male bias. Our hypothesis is that maybe males have an intrinsic islet fragility (perhaps from supporting a larger body mass), while females have an intrinsic susceptibility to autoimmune disease. In epidemiological terms, these two effects may cancel each other out, leading to similar levels of type 1 diabetes in males and females, but at a clinical level it may mean that different individuals would respond better to different treatment strategies.   

You identified two loci linked with insHEL-induced diabetes in the NOD^k mice.  What were your expectations about what you would find? Where you excited when Xrcc4 and Glis3 were identified as candidate genes?

Geneticists have been trying to work out the basis of spontaneous diabetes in NOD mice since the strain was first published in 1980. It turns out to be a very complex problem – there are more than 20 loci that contribute to diabetes susceptibility, and each time a loci is analysed in detail it ends up being a cluster of weaker loci working together. Decades later and we are only sure about a handful of candidates genes – so I didn’t have high expectations that we would progress far when looking at the genetics of insHEL-triggered diabetes. It turns out, however, that we had several major advantages. First, the genetics ended up being much simpler, with linkage only observed on two chromosomes. Second, because we knew which cell type was important – the beta cell – we were not operating in the dark about candidates. After filtering for expression in beta cells we were left with only a handful of candidates. Seeing Xrcc4 and Glis3 on the final list was bliss – they both made perfect biological sense. GLIS3 is one of the very few genes linked to both type 1 and type 2 diabetes in humans, and here we had it on our shortlist for a model that contains aspects of both diseases! It had taken more than 10 years to get to those two genes, but then we reached one of those dream runs in the laboratory where all the data just comes together, and every experiment gave support to the candidates.

You identify beta cell failure as a common link between T1D and T2D.  Are there ways that your findings can impact the clinical understanding or management of these diseases? 

In some ways, what we have here is the laboratory catching up to the clinic. The clinical overlaps between type 1 and type 2 diabetes have been apparent from the start, yet the research on genetics and animal models has consistently emphasized the differences. We may be in the process of reconciling these two approaches. The model that I favour is one where beta cell robustness or fragility lies at the centre of both diseases. In type 1 diabetes, failures in immune tolerance promote an attack on the beta cell, while in type 2 diabetes, hepatic insulin resistance leads to beta cell stress. In both cases, however, it may be the intrinsic robustness or fragility of the beta cell that dictate whether the pressure on beta cells remains subclinical or leads to diabetes. If this model holds true in patients then it would present a golden opportunity for preventing diabetes by increasing the robustness of beta cells.